Description
Describe the feature request.
@MOshima-PIFSC post about an error in the calculation of Mohn's Rho within the r4ss:SSmohsrho function has highlighted the benefit of streamlining code uses across organizations for diagnostics. I particularly liked how the ss3diag::SSplotRetro function output the yearly, combined, and the forecast Mohn's Rho which I think would be useful to all SS3 users. @kellijohnson-NOAA and I have had multiple discussions in the past about combining separate packages developed across Science Centers for model diagnostics. The two main packages that I am aware of are the ss3diags and the nwfscDiag packages and I think there is considerable overlap in what each of them do.
Describe alternatives you have considered
I think there are multiple pathways for streamlining efforts that I think we should discuss the pros and cons of each:
-
Incorporate all diagnostic functions with r4ss making them available to all users. For example, r4ss and the ss3diags packages both do calculations of Mohn's Rho using separate code. If we modified the r4ss::SSmohnsrho to output in an agreed upon format the same code could be called within separate packages yielding the same results.
-
Create a r4ss diagnostic package that does the core diagnostics (run single parameter profiles, jitters, retrospectives) currently available in the ss3diags and nwfscDiag packages with standardized plotting of results. This package could also contain additional diagnostics (ss3diags::SSplotRunstest) that are currently available in ss3diags that would be beneficial to all SS44 users.
-
Designate either ss3diag or nwfscDiag as the single diagnostic package and fold all functionality within both packages into one package. Alternatively, a new package could be created with each of the ss3diag or nwfscDiag package calling functions from the combined package that could provide backward compatibility (I think there are other variations of this for backward compatibility that we could also explore).
I suspect there are other options I have not thought of that also may be a viable pathway forward that achieves the goal of streamlining diagnostic functionality. I am open to all suggestions.
Statistical validity, if applicable
No response
Describe if this is needed for a management application
No response
Additional context
No response
Code of Conduct
- I agree to follow this project's Code of Conduct